---
name: eleven
description: Generate ElevenLabs-quality speech audio, paid per call in USDC on Base via x402. No API key, account, or subscription needed. Use when the user wants text-to-speech, voiceover, narration, or spoken audio from text.
---

# eleven, pay-per-call text-to-speech (tokens.fun)

Base URL: https://eleven.tokens.fun

## 1. Discover (free)

`GET https://eleven.tokens.fun/voices`

Returns available voices (each with a playable `preview_url`), TTS models, output
formats, defaults, and the pricing formula. No payment or auth required.

## 2. Speak (paid via x402)

`POST https://eleven.tokens.fun/speak` with JSON body:

```json
{ "text": "required, max 4000 chars", "voice_id": "optional", "model": "optional", "format": "optional", "speed": 1.0 }
```

- An unpaid request returns **HTTP 402** quoting the exact USDC amount for your text
  (x402 v2, scheme `exact`, network `eip155:8453`, USDC on Base).
- Pay with any x402 client (e.g. `@x402/fetch` wrapping fetch with a funded Base
  wallet) and retry. The same request body must be sent with the payment.
- Success returns `{ audio_url, characters, model, voice_id, format, content_type, size_bytes, charged_usd }`.
  The `audio_url` is a plain public download. Send `Accept: audio/mpeg` instead to
  stream raw audio bytes directly in the response.

## Pricing

Minimum $0.03 per call; $0.30/1K characters on flash/turbo
models, $0.60/1K on other models. The 402 always quotes the exact amount.

## Guarantees

- Payment settles **only after** synthesis succeeds. Failed calls are never charged.
- Each signed payment is single-use; sign a fresh payment per request.